Intelligent System Concepts
Intelligent System Concepts
System
Concepts
LEARNING OUTCOMES
By the end of this Topic, you should be able to:
1.
2.
3.
X INTRODUCTION
Since the introduction of computer system and data processing, human beings
have used both the computer system and data processing to make their day-today work easier. Therefore, work that usually consumes a lot of time requires
much computation and may be considered as trivial can be simplified and
automated. One limitation of computer technology is in the decision and
instruction given by a computer user. If a user makes a wrong decision and gives
instruction to the computer based on that wrong decision, then the outcome will
also be wrong. Due to this weakness, a field of study has been established to
enable machines and computers to think and make decisions in order to reduce
mistakes made by human beings. This field of study is called Artificial
Intelligence (AI).
1.1
Artificial intelligence (AI) is a field that studies methods or ways for machines
such as computers to act and do things like work and thinking similar to human
beings.
Most experts agree that artificial intelligence is closely connected to two main
ideas. First is to study the human thought processes to understand what
intelligence is all about. The second idea involves the representation of these
processes through machines such as computers and robots.
The abilities listed below represent intelligent behaviour or signs of intelligence:
1.2
DEFINITIONS OF ARTIFICIAL
INTELLIGENCE?
Figure 1.1: Categories of Artificial Intelligence Definition (Russell and Norvig, 1995)
Next, let us look at several definitions that can fit into each part of the figure
presented above.
1.2.1
Definitions under this category refer to cognitive modelling approach. Here, two
definitions are given as guidelines:
Definition 1:
'The exciting new effort to make computers think machines with
minds, in the full and literal sense'
(Haugeland, 1985)
Definition 2:
"[The automation of] activities that we associate with human
thinking, activities such as decision-making, problem solving,
learning"
(Bellman , 1978)
Cognitive modelling approach is based on cognitive science field of study.
Cognitive science combines artificial intelligence computer model with
experimental techniques from psychology field in an effort to build and test
theories regarding how the human mind functions.
If we were to claim that a particular system thinks like human beings, then we
must have ways to determine how they (human beings) think. There are two
ways to determine this. These are:
Psychological experiments.
If input, output and expression of time of a system are similar to a human being,
then this will be proof that there are mechanisms in the system that operate the
same way as in that particular human being. Thus, the fields of artificial
intelligence and cognitive science are beneficial to one another, particularly in the
study of sight, natural language and learning.
ACTIVITY 1.2
In your opinion, how can a machine think and act like a human
being, and then think and act rationally like a human being?
1.2.2
Definitions under this category refer to the Turing Test approach. Here, two
definitions are given as guidelines:
Definition 1:
"The art of creating machines that perform functions that require
intelligence when performed by people"
(Kurzweil, 1990)
Definition 2:
"The study of how to make computers do things at which, at the
moment, people are better"
(Rich and Knight, 1991)
Turing Test was introduced by Alan Turing (1950) to provide an acceptable and
workable definition of intelligence. Turing stated that an intelligent act is the
ability to achieve human level performance in all cognitive tasks that cannot be
distinguished by an interrogator. In summary, the test introduced by Turing is as
follows:
A computer is asked by a human interrogator through a teletype. The test is
successful when the interrogator is unable to determine whether a human or
computer answers the questions.
Turing test prevents physical interactions between the interrogator and computer
because physical stimulation of a person or an entity is not required to determine
intelligence. This test will be considered successful if a particular system requires
sophisticated functions such as natural language processing, knowledge
representation, automated reasoning, and machine learning. However, in
artificial intelligence, objective of an intelligent system to succeed in the Turing
test is not a priority in the decision-making process of a particular problem. It is
more focused on human and machine interactions and how a machine reacts to
instructions or stimulation by a person or user.
1.2.3
Definitions under this category refer to laws of thought approach. Here, two
definitions are given as guidelines:
Definition 1:
"The study of mental faculties through the use computational models"
(Charniak and McDermott, 1985
Definition 2:
"The study of the computations that makes it possible to perceive,
reason and act"
(Winston, 1992)
1.2.4
Definitions under this category refer to rational agent approach. Here, two
definitions are given as guidelines:
Definition 1:
"A field of study that seeks to explain and emulate intelligent
behaviour in terms of computational processes"
(Schalkoff, 1990
Definition 2:
"The branch of computer science that is concerned with the
automation of intelligent behaviour"
(Luger and Stubblefield, 1993
SELF-CHECK 1.1
What are the categories of definition for artificial intelligence?
State the differences between these categories.
1.3
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(i)
(ii)
(iii) Natural intelligence enables human beings to use their vast experience to
solve a problem
Artificial intelligence depends only on complete focus of one experience to
solve the same problem.
The advantages of natural intelligence in contrast to artificial intelligence show a
few limitations of artificial intelligence technology. However, there are a few
cases where artificial intelligence technology has the edge in terms of
productivity and quality. In order for a computer system to act more intelligently
than a human being, the system must be able to make the connection between
objects, events and processes easily like how an individual normally does
ACTIVITY 1.4
High expectation on artificial intelligence is the main problem faced in
the
beginning of the introduction of the technology. Why did it
happen? Is it still happening now?
ACTIVITY 1.5
The following is a list of a few day-to-day activities. A few of these
activities can be carried out by artificial intelligence, but there are also
a few that are impossible or difficult to be carried out. Mark those
activities that you feel suitable for artificial intelligence and those that
are suitable for natural intelligence.
Activity
Suitable for
1.
Recite poetry
2.
3.
4.
5.
Establish
between
countries
6.
7.
Obtain public
certain issue
8.
Debate on the
ancient texts
9.
diplomatic relations
two
neighbouring
support
on
meanings
a
of
11
1.4
ACTIVITY 1.6
After knowing the advantages of artificial intelligence and natural
intelligence, what do you think are the differences between artificial
intelligence and conventional computing?
The computing or processing and solving method of an intelligent system using
artificial intelligence technology is very different from conventional computing.
Here, the differences between the two methods are explained more clearly.
1.4.1
Conventional Computing
1.4.2
13
Even though problem solving in artificial intelligence is not directly done through
algorithm process, algorithm is still used in the search process. Artificial
intelligence is a unique approach for computer programming. Although artificial
intelligence is still a computer-based system, it has several distinctive
characteristics as shown in Table 1.1 below.
Table 1.1: Unique Characteristics of Artificial Intelligence Compared to
Conventional Computing
Dimension
Artificial Intelligence
Conventional
Computing
Processing
Usually symbolic
Usually algorithm
Input form
Can be incomplete
Must be complete
Search
Use algorithm
Explanation
Provided
Usually not
provided
Focus
Knowledge
Data, information
Structure
Control is integrated
with data
Output form
Can be incomplete
Must be complete
Easy since
modular
Hardware
All types
Determination power
None
it
is
very
Usually difficult
ACTIVITY 1.7
Sometimes we can be confused with what differentiates a system based
on artificial intelligence technology from a system that functions using
conventional computing. The table below contains several statements
from software, device, or system owners, and all said that their systems
use artificial intelligence. Your task is to identify which one really uses
artificial intelligence.
Statement
Yes (
No ( )
Yes (
No ( )
Yes (
No ( )
Yes (
No ( )
Yes (
No ( )
1.5
15
1.5.1
Expert System
1.5.2
1.5.3
Speech Recognition
1.5.4
Robotics
When artificial intelligence is combined with sensory system, sight system, touch
system, and signal processing system, it falls under a category called robotics. An
intelligent robot has a sensory tool such as a camera to gather information related
to operations and to the robotic environment. The intelligent part of the robot
allows it to interpret the information gathered, to act and adapt to the
environment, and not follow instructions only.
ACTIVITY 1.8
(a)
(b)
(c)
In Topic 1, you have learnt about artificial intelligence, its definitions, and
several examples of artificial intelligence applications. You have also seen the
differences between artificial intelligence technology and other current
technologies. You should be able to understand:
(i)
(ii)
(iv)
x